6938
tinct advantage over the X-ray and neutron diffraction techniques. This is particularly the case for the investigation of materials with characteristic local motion. The sensitivity of the ednmr technique to the determination of the relative positions of light atoms such as protons makes this technique highly complementary to the X-ray crystallographic technique. (It should be noted that the molecular structure, as distinct from the crystal structure, is also partially accessible to nmr studies carried out in nematic solvents.40) Ednmr and Hydrogen Bonding. The present studies and investigation of malonic acid by Kwiram and McCalley18s22indicate that the empirical equation of Soda and Chiba41
(40) See, for example, the results for cyclobutane reported by S . Meiboom and L. C. Snyder, J . Chem. Phys., 52, 3857 (1970). (41) G. Soda and T.Chiba, ibid., 50,439 (1969).
where R ( O . . * H )is given in Bngstrom units, should be modified for application to hydrogen-bonded carboxylic acid bridge structures. A more appropriate relationship between the quadrupole coupling constants and the length of the hydrogen bond appears to be
although we emphasize that further studies are necessary to refine the coefficients. Comparison of the ednmr data for 1,l-CBDA with that for malonic acid1B,22indicates that somewhat stronger hydrogen bonding exists in 1,l-CBDA. This difference in hydrogen bond strengths is also reflected in the melting points: 157" (1,l-CBDA) and 135.6" (malonic acid). Acknowledgments. The authors express their appreciation to Dr. R. C. McCalley for many stimulating and informative discussions.
Phenalenyl Dimer Cation and Its Electron Paramagnetic Resonance Spectrum'" D. H . Paskovichlb and A. H . Reddoch" Contribution f r o m the Division of Chemistry, National Research Council of Canada, Ottawa K I A OR6. Received January 31, 1972
Abstract: A new type of dimer cation, formed from a neutral radical and its diamagnetic cation, is reported here. The phenalenyl dimer cation (CI3H9)2+is formed on dissolving phenalene in CF3COCF3.nHs0,n = 2-5. Its epr spectrum is described by the coupling constants a1 (12 H) = 3.108 and a2 (6 H) = 0.920 G and a g value of 2.002721.
T
he epr spectrum of the neutral phenalenyl radical was first reported by Calvin, et U I . , ~who found it in a CC14 solution containing phenalene which had been exposed to air. This radical can be formed in a wide range of organic solvents by dissolving phenalene, which reacts with atmospheric oxygen presumably forming a hydroperoxide, which on decomposition leads to the radical. When the solution is degassed, a well-resolved epr spectrum is obtained. The coupling constants and g value are slightly solvent dependent, but typical values are al = 6.31 G, a? = 1.81 G, and g = 2.00260.3-h A different spectrum is obtained when a range of hexafluoroacetone (HFA) hydrates, CF3COCF3nHaO, n = 2-5, is used as a solvent. This spectrum consists of 13 septets. For n = 4.2 the coupling constants are al (12 H) = 3.108 i 0.002 G, a? (6 H) = 0.920 f 0.002 G, and the g value is 2.002721 f 0.000005, the latter (1) (a) Issued as NRCC No. 12688. (b) National Researchcouncil of Canada Visiting Scientist 1967-1968. (2) P. E. Sogo, M. Nakazaki, and M. Calvin, J . Chem. Phjs., 26, 1343 (1957). (3) A . H. Reddoch, D . H. Paskovich, and A. E. Goggins, to be submitted for publication. (4) J. E. Bennett, Proc. Chem. Soc., 144 (1961). (5) E. G. Segal, M. Kaplan, and G. IC. Fraenkel, J . Chem. P h ~ s .43, , 4191 (1965). (6) A . H. Reddoch, C. L. Dodson, and D. H. Paskovich, ibid., 52, 2318 (1970).
Journal of the American Clremical Society 1 94:20
having been corrected for second-order effects.' N o signal was obtained with n > 6, and usually one was not obtained if M 'v 1.5, although results were not completely reproducible in this nearly solid hydrate. When dimethoxyethane was present as well as H F A . 1.5H20only the phenalenyl signal was obtained. When DzO was used instead of HzO no deuteration of either phenalenyl or the new radical was detected in their spectra. In one experiment in HFA.2.5Dz0 the spectra of phenalenyl and the new species were superimposed : however, the former species disappeared in a short time. Precise absolute coupling constants and g values were not obtained for this sample but good differential measurements were obtained and will be discussed below. Since the new spectrum may be derived from that of phenalenyl by doubling the number of each type of proton and halving the coupling constants, it is clearly a dimeric species. Since the spectra are quite sharp. (the dimer line width is 88 mG compared with 80 mG for monomer in the same sample) there can be only one unpaired electron, pointing to a dimeric ion. The chemistry indicates that this must be the dimer cation (C13H&+. Thus the anion would be unlikely to form (7) R . W. Fessenden, ibid., 37, 747 (1962).
October 4, 1972
6939
in an acid such as the HFA hydrate.s The cation might conceivably be formed directly from the phenalene, since HFA is reportedgl1Oto be a hydride ion acceptor. However a more likely route is suggested by the presence of phenalane C13H12, found in the reaction mixture by thin layer chromatography. Thus phenalene and the related 1,2,3,5- and l,2,3,6-tetrahydropyrenes1l protonate in strong acid and then undergo intermolecular hydride transfer to yield the corresponding phenalenium cations and phenalanes. A similar disproportionation reaction seems to occur in the present system. The new species is thus analogous to a number of dimer cations which have been studied by optica112j13 and epr methods including those derived from naphthalene, l 4 anthracene, l 4 pyrene,16 perylene, l 6 coronene,'7 and mesitylene.l8 The analogy may not be complete, however, for while these dimers have lost one electron from two closed shell molecules, the phenalenyl dimer cation has lost an electron from two open shell molecules. The latter system thus has one electron in the two orbitals derived from the two nonbonding orbitals of its two components, while the other dimers have three electrons in the pair of orbitals based on the highest bonding orbitals of their components. Thus the phenalenyl dimer cation is a new class of dimer cation. Several dimer cations have been used to test14 the validity of the Colpa and Bolton relation19s20 and that of Giacometti, et ~ 1 1 . (A ~ ~discussion of these relations (8) C. G. Krespan and W. J. Middleton, Fluorine Chem. Reu., 1, 145 (1967). (9) N. P. Gambaryan and I. L. Knunyants, I Z L . Akad. Nauk SSSR, Ser. Khim., 4, 722 (1965). (10) A. F. Janzen and C. J. Willis, Znorg. Chem., 6, 1900 (1967). (11) D. H. Paskovich and N. C. Das, Chem. Commun., 39 (1967). (12) B. Badger and B. Brocklehurst, Trans. Faraday Soc., 66, 2939 ( 1970). (13) A. Kira, S. Arai, and M. Imamura, J . Chem. Phys., 54, 4890 (1971). (14) 0. W. Howarth and G. K . Fraenkel, ibid., 52, 6258 (1971). (15) J. T. Cooper and W. F. Forbes, Can. J . Chem., 46, 1158 (1968). (16) T. C. Chiang and A. H. Reddoch, J. Chem. Phys., 52, 1371 ( 1970). (17) H. van Willigen, E. de Boer, J. T. Cooper, and W. F. Forbes, ibid., 49, 1190 (1968). (18) S. Shih and R. M. Dessau, ibid., 55, 3757 (1971). (19) J. P. Colpa and J. R. Bolton, Mol. Phys., 6,273 (1963). (20) J. R . Bolton, J . Chem. Phys., 43, 309 (1965). (21) G. Giacometti, P. L. Nordio, and M. V. Pavan, Theor. Chim. Acta, 1, 404 (1963).
has recently been given by Vincow.22) However, it is known that proton coupling constants can vary by a few per cent because of vibrational6and solvent3effects. Since the changes in coupling constants predicted by the above relations are not much greater than these environmental effects, a convincing test would require the elimination of these latter effects. At present this cannot be readily done, particularly for the dimer cations. The ratio of the coupling constants of the phenalenyl monomer (alf) and dimer cation (UD)were measured from the spectrum of the sample containing both species. This procedure does not necessarily avoid the environmental effects. The results are allT/aID = 2(1.027 f 0.001) and a2M/a2D = 2(1.001 * 0.001). This result is not consistent with the relation of Giacometti, et al., which would require the two ratios to be equal. The Colpa-Bolton relation predicts that a 2 ~ / aZD= 2, since the charge densities at these positions are zero; however, it predicts that allf/alD < 2, since there is a positive charge at alD of e/12. Thus neither relation is supported by the data but the environmental effects may have a considerable influence. Similar difficulties may affect the interpretation of g values. From the same spectrum used above gD g, = (9.1 f 0.1) X 10-5 (corrected7 value 9.6 X 10-j). This may be compared with perylene16 where g D (- 80") - gbf(25") = -(9.2 0.9) X 10-j in CH2C12and g D (-20") - g,(25") = -(I f 2) X 10-j in CH3CN. For coronene17gD - g M = (1 f 3) X 10-j using different solvents for monomer and dimer. Clearly discussion is premature. As with the other dimer cations, the retention of the equivalence of the protons of the monomer components suggests that the dimer cations have a sandwich type structure, although it does not prove it since the spectrum could be merely the time average over a motion through a set of positons of lower symmetry. This structure is quite different from that postulated23 $ 2 for the dimer which is in equilibrium with the phenalenyl radical. The latter has been regarded as a ubonded dimer. The present dimer cation because of its symmetry cannot have such a structure.
*
(22) G. Vincow, J . Phys. Chem., 75, 3400 (1971). (23) D . H. Reid, Chem. Ind. (London), 1504 (1956). (24) F. Gerson, E. Heilbronner, A. H. Reddoch, D. H. Paskovich, and N. C. Das, Helu. Chim. Acta, 50, 813 (1967).
Paskovich, Reddoch 1 Phenalenyl Dimer Cation